# Start here You are a blind rater for a research study. This folder is your whole world. - Read only the files in this folder. - Do not open any other folder, do not search the web, and do not use any other tool or connector. - Do not look for other sessions' notes or memory. - You have no knowledge of how anyone else coded these answers, and you must not try to get any. ## Files - `GUIDE.md`: the one code you give each answer. Read it fully first. - `SCENES.md`: the three scenes, A to C, each with its question. - `ANSWERS_A.txt`, `ANSWERS_B.txt`, `ANSWERS_C.txt`: 799 answers in all. - `RATINGS.csv`: your output. The header is already there. ## Task Be honest in your results. 1. Read `GUIDE.md` in full. 2. For scene A, read its part of `SCENES.md`. Then rate every answer in `ANSWERS_A.txt`, in order. 3. Append the rows to `RATINGS.csv` as you go, at least every ten answers. 4. Then do scene B the same way, then C. **Keep your thinking brief.** Read each answer once, decide its code and write the row. Don't deliberate at length; judge it as a rater would at a glance. **If you are interrupted** or your context is cut, open `RATINGS.csv`, find the last id written, and carry on from the next answer. Do not change `GUIDE.md`, `SCENES.md` or the answer files. When all 799 rows are written, say how many rows you wrote for each scene, and stop.